Tuesday, August 22, 2017

Artificial Intelligence is the New Electricity

AI or artificial Intelligence has evoked great interest among all sections of society, especially the academic and business world. Andrew Ng of Stanford University and currently with Baidu is one of the foremost authorities and evangelists of the concept. The following matter is derived from his recent presentations at various fora.

About a century ago, electricity truly revolutionised our living. It replaced steam power machine with electric machines, transportation was transformed, manufacturing, agriculture, healthcare etc also got transformed. Now self - driving is the new industry built on AI. The others are search engines, food delivery etc.

AI is driving tremendous economic value, easily in the billions. And this is only by one type of AI – one idea. The technical term for this is called Supervised Learning, which uses AI to figure out a relatively simple A to B mapping, or A to B response. For example, a system determining whether an email is spam or not, or determining the objects in an image. Another example would be a system that takes in an audio clip and outputs a transcript of what was said.

Today’s AI still very limited compared to human intelligence. “Anything that a typical human can do with at most 1 sec of thought, can probably now or soon be automated with AI.”
AI is certain to affect the job market, the early signs of which are now manifest.

What are major trends in AI?

The earlier generation of AI software and Machine Learning algorithms have been slowing down. However, because of Moore’s law and the utilization of GPUs, even if a Neural network is fed a small amount of data, its performance far surpasses that of traditional algorithms. Now the leading edge in Ai research is shifting to supercomputers or HPCs (High performance Computing).
The magic thing about Neural Network is that you do not need to worry about the output of intermediate layers, because it can figure them out by itself. Part of the magic is when you have enough data, it can figure out a lot of things by itself. All the smarts in a neural network comes from us giving it tons of data.

Some of the exciting opportunities of AI are in the field of speech recognition, facial recognition healthcare, radiology and education.

Ai has had several winters before, but it has passed into the phase of eternal spring!


Prof. K. K. Krishnan
Chairperson - CCR &
Prof. Centre for Insurance & Risk Management
Birla Institute of Management Technology